This stack is designed to allow you to monitor your AWS Lambda errors and send alerts to a specified slack channel. There are plenty of paid services (and even some free ones) that offer the same functionality, but the largest drawback is that it requires you to send your logs to a vendor's AWS account. This solution allows you to keep the logs in your account and customize which Lambda's forward their logs to slack.
The current implementation picks up and Lambda Error, Timeout, and any message written with console.error()
.
The subscriber lambda is invoked any time you update tags on a Lambda function. The subscriber looks for a tag named monitoring
and looks for that tag value to evaluate to true
. If the subscriber finds the monitoring
tag is set to true
then it creates a new log subscription on the updated lambda. This log subscription subscribes the Alerter
lambda to the modified lambda's log group.
If the monitoring tag does not exist or is set to false, this Lambda will check if the Alerter
function is subscribed to the log groups and remove the subscription if so. This allows you to toggle on and off the subscriptions as needed.
This lambda is invoked by Cloudwatch Logs when a lambda with the monitoring
tag writes logs. This lambda then parses the log event to find out if it is a function error or a console.error()
. It then relays this information to a specified slack channel.
At this time, this app requires you to specify a Slack Webhook URL. To generate a Slack Webhook URL, follow the Create The Slack App
portion of this Medium post. Once you have a Webhook URL, you need to add it to the serverless.yml
file in the alerter
function's environment
settings. If your Slack Webhook URL looks like this:
https://hooks.slack.com/services/ABCDEF123/12345ABCDEF/XXXXYYYYXXXXYYYYXXXXYYYY
then update the environment variable to look like this:
alerter:
handler: alerter/index.main
environment:
SLACK_WEBHOOK_URL: ABCDEF123/12345ABCDEF/XXXXYYYYXXXXYYYYXXXXYYYY